Echo State Incremental Gaussian Mixture Network for Spatio-Temporal Pattern Processing

Abstract. This work introduces a novel neural network algorithm for online spatio-temporal pattern processing, called Echo State Incremental Gaussian Mixture Network (ESIGMN). The proposed algorithm is a hybrid of two state-of-the-art algorithms: the Echo State Network (ESN), used for spatio-temporal pattern processing, and the Incremental Gaussian Mixture Network (IGMN), ap-plied to aggressive learning in online tasks. The algorithm is compared against the conventional ESN in order to highlight the advantages of the IGMN ap-proach as a supervised output layer. Resumo. Este trabalho introduz um novo algoritmo de redes neurais para processamento online de padrões espaço-temporais, chamado Echo State In-cremental Gaussian Mixture Network (ESIGMN). O algoritmo proposto e ́ um hı́brido de dois algoritmos estado-da-arte: a Echo State Network (ESN), usada para processamento de padrões espaço-temporais, e a Incremental Gaussian Mixture Network (IGMN), aplicada ao aprendizado agressivo em tarefas on-line. O algoritmo e ́ comparado com a ESN convencional a fim de destacar as vantagens da abordagem IGMN como camada supervisionada de saı́da. 1